Instructions to use kfkas/Llama-2-ko-7b-Chat with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use kfkas/Llama-2-ko-7b-Chat with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="kfkas/Llama-2-ko-7b-Chat")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("kfkas/Llama-2-ko-7b-Chat")
model = AutoModelForCausalLM.from_pretrained("kfkas/Llama-2-ko-7b-Chat", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use kfkas/Llama-2-ko-7b-Chat with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "kfkas/Llama-2-ko-7b-Chat"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "kfkas/Llama-2-ko-7b-Chat",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/kfkas/Llama-2-ko-7b-Chat

SGLang

How to use kfkas/Llama-2-ko-7b-Chat with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "kfkas/Llama-2-ko-7b-Chat" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "kfkas/Llama-2-ko-7b-Chat",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "kfkas/Llama-2-ko-7b-Chat" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "kfkas/Llama-2-ko-7b-Chat",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use kfkas/Llama-2-ko-7b-Chat with Docker Model Runner:
```
docker model run hf.co/kfkas/Llama-2-ko-7b-Chat
```

Prompt 문의

by goreng2 - opened Oct 26, 2023

Discussion

goreng2

Oct 26, 2023

Chat 버전으로 Finetune 하실 때, Prompt format을 어떻게 하셨는지 궁금합니다.

kfkas

Owner Oct 26, 2023

KuLLM의 포맷을 사용하였습니다

goreng2

Oct 26, 2023

•

edited Oct 26, 2023

Prompt template라고 표현하는게 더 맞는 것 같습니다.

Llama2-chat 모델의 경우, 아래와 같은 서식을 사용하는 것으로 알고 있습니다. (https://gpus.llm-utils.org/llama-2-prompt-template/)

<s>[INST] <<SYS>>
{your_system_message}
<</SYS>>

{user_message_1} [/INST]

KuLLM의 포맷을 사용하셨다는 말은 위 서식의 {your_system_message}에 Instruction 열 내용을, {user_message_1}에 Input 열 내용을 대입하셨다는 말씀이실까요?

kfkas

Owner Oct 26, 2023

{
"description": "Alpaca-LoRA에서 사용하는 템플릿입니다.",
"prompt_input": "아래는 작업을 설명하는 명령어와 추가 컨텍스트를 제공하는 입력이 짝을 이루는 예제입니다. 요청을 적절히 완료하는 응답을 작성하세요.\n\n### 명령어:\n{instruction}\n\n### 입력:\n{input}\n\n### 응답:\n",
"prompt_no_input": "아래는 작업을 설명하는 명령어입니다. 요청을 적절히 완료하는 응답을 작성하세요.\n\n### 명령어:\n{instruction}\n\n### 응답:\n",
"response_split": "### 응답:"
}
이런식입니다.

goreng2

Oct 27, 2023

감사합니다!

goreng2 changed discussion status to closed Oct 27, 2023

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment